Accuracy Of Automated Measurement Of Angles Of Strabismus With DAISY In Relation To Direction Of Gaze And Free Head Rotation

Abstract

Purpose: : The Delft Assessment Instrument for Strabismus in Young children (DAISY) aims to measure angles of strabismus accurately in young children to improve outcome of surgery.

Methods: : Stereo images were obtained with two Prosilica GC2450 cameras to derive angles of strabismus, using the coordinates of the pupil centers, fixation objects (LEDs) and corneal reflections created by three infrared light sources. Head rotation was assessed with a third camera and a face-tracking system (FaceAPI). Three types of validation experiments were performed on orthotropic test subjects at a measurement distance of 1.10 meter in roomlight. I. The error in measurement of strabismus angle was determined in one adult with a fixed head pose. II. The error in measurement of head rotation was determined with a high-precision secondary measurement system in ten adults and ten children. III. The error in measurement of strabismus angle with free head movement was determined in three adults. Gaze range was 50 degrees horizontally and vertically.

Results: : I. Error of strabismus angle was <0.5° up to 15° and <0.7° for up to 25° eccentric gaze. II. The error of the head rotation estimation with the FaceAPI was <5° for 99.5% of the measurements in adults and for 99.2% of those in children’s. The median of the error of the adults' data was 1.40° and was 1.23° for the children's data. III. The overall error of horizontal angle of strabismus with unrestrained head was 0.34° for gaze ahead and 0.53° for eccentric gaze. The runtime for a single measurement was 2ms.

Conclusions: : Variance introduced by inaccurate measurement with DAISY is clinically not relevant.

Keywords: strabismus • eye movements • strabismus: diagnosis and detection